Missing SECTION break

Andreas Mengel · New User Joined: 26 Nov 2019 Posts: 4 Location: Germany

Hi,

I have a strange problem building a report from two datasets. I want to have a blank line in output every time a specific part of my input changes. This is working quite fine in 99% of the cases, but sometimes just one of the section breaks does not occur and it seems to depend on the input data:
- for some input files all the breaks are there
- for some input files just ONE break is missing, all others are as expected
- different input files show the problem at different lines

Sorry if my testdata is a bit lengthy, but the problem does only occur with specific input data.

In the case of files SRTFNF1A and SRTJNF2A the missing break is between line 304 and 305.

SRTJNF1B and SRTJNF2B contain just the data for the above sections with the missing break and the one before and after it in the same order as in SRTJNF1A/2A. In this case the output is as expected.

SRTJNF1* should be FB 70
SRTJNF2* should be FB 53

Best regards
Andreas

Andreas Mengel · New User Joined: 26 Nov 2019 Posts: 4 Location: Germany

Hi,

I forgot to add the SDSF output from the job...

Regards,
Andreas

sergeyken · Posted: Wed Nov 27, 2019 6:13 pm

1. Include your samples within code tags, without attachments.

2. Limit your test data to only that part causing this problem, and then provide intermediate results from SORT, as detailed as possible. Usually I start from the most primitive SORT statements, and then add extra processing, step by step, each time verifying details of intermediate results.

This must be standard approach for ANY SOFTWARE CODE DESIGN, AND DEBUGGING!
One should clearly understand this before selecting software engineering job for himself.

You cannot first build a skyscraper, and at the end test: how strong is its basement?

Andreas Mengel · New User Joined: 26 Nov 2019 Posts: 4 Location: Germany

Hi sergeyken,

sergeyken · Posted: Wed Nov 27, 2019 9:44 pm

Please, limit your sample to only the part causing the problem, and post the BAD data with BAD result only; also optionally - limited GOOD data with GOOD result.

The most of readers would not do this part of job for you; but some of them would, maybe.

sergeyken · Posted: Wed Nov 27, 2019 9:50 pm

I suspect that SKIP=1L results in adding CC='-' as print control character, which is successfully removed by REMOVECC parameter. I can only guess, without presented intermediate results...

I recommend to try using HEADER3=(/) instead.

sergeyken · Posted: Wed Nov 27, 2019 10:06 pm

Joerg.Findeisen · Posted: Thu Nov 28, 2019 12:48 am

@sergeyken: Your conclusion sounds logical. The SKIP parm works with pages, HEADER works with data. Need to try myself at some point.

Andreas Mengel · New User Joined: 26 Nov 2019 Posts: 4 Location: Germany

Hi again and thanks for the reminder to RTFM. Should have RTFMed more closely

SKIP=xL does in fact add x blank lines to the report instead of print control characters. The problem with my report was that SKIP does only add the desired empty lines IF the following section will also start on the same page as the previous section. I misread the manual on that. So in my case the missing break was after line 304 which is in fact right at the end of page 6 counting that each page has 60 lines and the report header has 4 lines...

So the solution for me is to simply use TRAILER3=(X) instead which will always add an empty line following any section regardless of any page breaks.

sergeyken · Posted: Mon Dec 02, 2019 8:30 pm

That's why from the very beginning I suggested first of all to verify intermediate results from a limited test data. All these issues would become clear immediately.

For myself it is obvious approach; I use it all my life.